Year: 2025
Type: conference-paper
Abstract: The financial industry faces a substantial workload in verifying document images.Existing methods based on visual features struggle to identify fraudulent document images due to the lack of visual clues on the tampering region.This paper proposes CSIAD (Cross-Sample Image Anomaly Detection) by leveraging LLMs to identify logical inconsistencies in similar images.This novel framework accurately detects forged images with slight tampering traces and explains anomaly detection results.Furthermore, we introduce CrossCred, a new benchmark of real-world fraudulent images with fine-grained manual annotations.Experiments demonstrate that CSIAD outperforms state-of-the-art image fraud detection methods by 79.6% (F 1) on CrossCred and deployed industrial solutions by 21.7% (F 1) on business data.The benchmark is available at https://github.com/XMUDM/CSIAD.
Authors: QiWen Wang, Junqi Yang, Zhenghao Lin, Zhenzhe Ying, Weiqiang Wang, Chen Lin
Institutions: Xiamen University, Ant Group (China)
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4412945221